In this paper, the development of a multivariable control system based on Model Predictive Control (MPC) is described and experimentally validated for a diesel engine with a complex dual-loop Exhaust Gas Recirculation (EGR) and Variable Geometry Turbocharger (VGT) air handling system. This air architecture is required to meet the stringent emissions and fuel economy requirements for this engine. The air system is a multi-input multi-output nonlinear system that makes the design and tuning of the controller challenging. In this paper, a switched linear MPC with feed-forward and feedback control architecture is developed for the system. The presented design process is systematic, modular, and model-based leading to a considerable reduction in the development and calibration time of the control system. A significant improvement is also observed with respect to a well-tuned feed-forward controller. The design process and the performance of the controller are demonstrated experimentally in the engine test cell using Matlab/SIMULINK software and dSPACE rapid prototyping tool.
展开▼
